1
|
Goyal M, Tafe LJ, Feng JX, Muller KE, Hondelink L, Bentz JL, Hassanpour S. Deep Learning for Grading Endometrial Cancer. THE AMERICAN JOURNAL OF PATHOLOGY 2024:S0002-9440(24)00202-5. [PMID: 38879079 DOI: 10.1016/j.ajpath.2024.05.003] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Track Full Text] [Subscribe] [Scholar Register] [Received: 11/27/2023] [Revised: 05/10/2024] [Accepted: 05/17/2024] [Indexed: 06/26/2024]
Abstract
Endometrial cancer is the fourth most common cancer in women in the United States; the lifetime risk for developing this disease is approximately 2.8%. Precise histologic evaluation and molecular classification of endometrial cancer are important for effective patient management and determining the best treatment modalities. This study introduces EndoNet, which uses convolutional neural networks for extracting histologic features and a vision transformer for aggregating these features and classifying slides based on their visual characteristics into high- and low-grade cases. The model was trained on 929 digitized hematoxylin and eosin-stained whole-slide images of endometrial cancer from hysterectomy cases at Dartmouth-Health. It classifies these slides into low-grade (endometrioid grades 1 and 2) and high-grade (endometrioid carcinoma International Federation of Gynecology and Obstetrics grade 3, uterine serous carcinoma, or carcinosarcoma) categories. EndoNet was evaluated on an internal test set of 110 patients and an external test set of 100 patients from The Cancer Genome Atlas public database. The model achieved a weighted average F1 score of 0.91 (95% CI, 0.86 to 0.95) and an area under the curve of 0.95 (95% CI, 0.89 to 0.99) on the internal test, and 0.86 (95% CI, 0.80 to 0.94) for F1 score and 0.86 (95% CI, 0.75 to 0.93) for area under the curve on the external test. Pending further validation, EndoNet has the potential to support pathologists without the need of manual annotations in classifying the grades of gynecologic pathology tumors.
Collapse
Affiliation(s)
- Manu Goyal
- Department of Biomedical Data Science, Dartmouth College, Hanover, New Hampshire.
| | - Laura J Tafe
- Department of Pathology and Laboratory Medicine, Dartmouth-Health, Lebanon, New Hampshire
| | - James X Feng
- Geisel School of Medicine, Dartmouth College, Hanover, New Hampshire
| | - Kristen E Muller
- Department of Pathology and Laboratory Medicine, Dartmouth-Health, Lebanon, New Hampshire
| | - Liesbeth Hondelink
- Department of Pathology, Leiden University Medical Center, Leiden, the Netherlands
| | - Jessica L Bentz
- Department of Pathology and Laboratory Medicine, Dartmouth-Health, Lebanon, New Hampshire
| | - Saeed Hassanpour
- Department of Biomedical Data Science, Dartmouth College, Hanover, New Hampshire
| |
Collapse
|
2
|
Alwafai Z, Beck MH, Fazeli S, Gürtler K, Kunz C, Singhartinger J, Trojnarska D, Zocholl D, Krankenberg DJ, Blohmer JU, Sehouli J, Pietzner K. Accuracy of endometrial sampling in the diagnosis of endometrial cancer: a multicenter retrospective analysis of the JAGO-NOGGO. BMC Cancer 2024; 24:380. [PMID: 38528468 PMCID: PMC10964509 DOI: 10.1186/s12885-024-12127-7] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/27/2023] [Accepted: 03/15/2024] [Indexed: 03/27/2024] Open
Abstract
BACKGROUND Accurate preoperative molecular and histological risk stratification is essential for effective treatment planning in endometrial cancer. However, inconsistencies between pre- and postoperative tumor histology have been reported in previous studies. To address this issue and identify risk factors related to inaccurate histologic diagnosis after preoperative endometrial evaluation, we conducted this retrospective analysis. METHODS We conducted a retrospective analysis involving 375 patients treated for primary endometrial cancer in five different gynaecological departments in Germany. Histological assessments of curettage and hysterectomy specimens were collected and evaluated. RESULTS Preoperative histologic subtype was confirmed in 89.5% of cases and preoperative tumor grading in 75.2% of cases. Higher rates of histologic subtype variations (36.84%) were observed for non-endometrioid carcinomas. Non-endometrioid (OR 4.41) histology and high-grade (OR 8.37) carcinomas were identified as predictors of diverging histologic subtypes, while intermediate (OR 5.04) and high grading (OR 3.94) predicted diverging tumor grading. CONCLUSION When planning therapy for endometrial cancer, the limited accuracy of endometrial sampling, especially in case of non-endometrioid histology or high tumor grading, should be carefully considered.
Collapse
Affiliation(s)
- Zaher Alwafai
- Department of Gynecology and Obstetrics, University of Greifswald, Greifswald, Germany
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
| | - Maximilian Heinz Beck
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany.
- Department of Gynecology With Center for Oncological Surgery, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Virchow Klinikum, Berlin, Germany.
- Department of Gynecology With Breast Center, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Mitte, Berlin, Germany.
| | - Sepideh Fazeli
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Klinik Für Gynäkologie, Krankenhaus Waldfriede, Berlin, Germany
| | - Kathleen Gürtler
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Klinik Für Gynäkologie, DRK-Kliniken Berlin Westend, Berlin, Germany
| | - Christine Kunz
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Department of Gynecology and Obstetrics, Krankenhaus St. Elisabeth Und Barbara, Halle, Germany
| | - Juliane Singhartinger
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Department of Gynecology and Obstetrics, Klinikum Traunstein, Traunstein, Germany
| | - Dominika Trojnarska
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Faculty of Health Sciences, Jagiellonian University Medical College, Cracow, Poland
| | - Dario Zocholl
- Institute of Biometry and Clinical Epidemiology, Charité - Universitätsmedizin Berlin, corporate member of, Freie Universität Berlin and Humboldt-Universität Zu Berlin, Berlin, Germany
| | - David Johannes Krankenberg
- Department of Gynecology With Breast Center, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Mitte, Berlin, Germany
| | - Jens-Uwe Blohmer
- Department of Gynecology With Breast Center, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Mitte, Berlin, Germany
| | - Jalid Sehouli
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Department of Gynecology With Center for Oncological Surgery, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Virchow Klinikum, Berlin, Germany
| | - Klaus Pietzner
- Young Academy of Gynecologic Oncology (JAGO), Berlin, Germany
- Department of Gynecology With Center for Oncological Surgery, Charité-Universitätsmedizin Berlin, corporate member of Freie Universität Berlin, Humboldt-Universität Zu Berlin, Berlin Institute of Health, Charité Universitätsmedizin Berlin, Campus Virchow Klinikum, Berlin, Germany
| |
Collapse
|
3
|
Rao Q, Liao J, Li Y, Zhang X, Xu G, Zhu C, Tian S, Chen Q, Zhou H, Zhang B. Application of NGS molecular classification in the diagnosis of endometrial carcinoma: A supplement to traditional pathological diagnosis. Cancer Med 2023; 12:5409-5419. [PMID: 36341543 PMCID: PMC10028062 DOI: 10.1002/cam4.5363] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/10/2022] [Revised: 09/26/2022] [Accepted: 10/05/2022] [Indexed: 11/09/2022] Open
Abstract
OBJECTIVE This study aims to demonstrate the advantages of NGS molecular classification in EC diagnosis and to assess whether molecular classification could be performed on curettage specimens and its concordance with subsequent hysterectomy specimens. METHODS 80 patients with hysterectomy specimens and 35/80 patients with paired curettage specimens were stratified as POLE mut, MSI-H, TP53 wt, or TP53 abn group by NGS panel. Histotype, tumor grade, IHC results, and other pathological details were taken from original pathological reports. RESULTS The correlation analysis of 80 patients with hysterectomy specimens between NGS molecular classification and clinicopathological characters displayed that the POLE mut group was associated with EEC (87.5%) and TP53 abn subtype was correlated to a later stage (Stage II-IV, 47.6%), G3 (76.2%), serous histology (61.9%) and myometrial invasion ≥50% (47.6%). A favorable concordance (31/32, 96.9%) was shown in MSI analysis and MMR IHC results, and the agreement rate of p53 IHC and TP53 mutation was 81.5% (53/65). Compared with the p53 IHC abnormal group, the TP53 mutation group had a higher correlation with high-risk factors. A high level of concordance (31/35, 88.0%) of NGS molecular classification was achieved between curettage specimens and hysterectomy specimens while grade and histotype (including unclassified group) from curettage specimens and hysterectomy specimens showed only moderate levels of agreement, 54.3% (19/35) and 68.6% (24/35), respectively. CONCLUSION NGS molecular classification achieved on curettage samples showed high concordance with the final hysterectomy specimens, demonstrating superior to the conventional pathological assessment of grade and histotype and potential utilization in clinical practice.
Collapse
Affiliation(s)
- Qunxian Rao
- Department of Gynecologic Oncology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China
| | - Jianwei Liao
- Cellular and Molecular Diagnostics Center, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China
| | - Yangyang Li
- Department of Pathology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China
| | - Xin Zhang
- Department of Ultrasound, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, China
| | - Guocai Xu
- Department of Gynecologic Oncology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China
| | - Changbin Zhu
- Department of Translational Medicine, Amoy Diagnostics Co., Ltd., Xiamen, China
| | - Shengya Tian
- Department of Translational Medicine, Amoy Diagnostics Co., Ltd., Xiamen, China
| | - Qiuhong Chen
- Department of Translational Medicine, Amoy Diagnostics Co., Ltd., Xiamen, China
| | - Hui Zhou
- Department of Gynecologic Oncology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China
| | - Bingzhong Zhang
- Department of Gynecologic Oncology, Sun Yat-sen Memorial Hospital, Sun Yat-sen University, Guangzhou, People's Republic of China
| |
Collapse
|
4
|
Kasius JC, Pijnenborg JMA, Lindemann K, Forsse D, van Zwol J, Kristensen GB, Krakstad C, Werner HMJ, Amant F. Risk Stratification of Endometrial Cancer Patients: FIGO Stage, Biomarkers and Molecular Classification. Cancers (Basel) 2021; 13:cancers13225848. [PMID: 34831000 PMCID: PMC8616052 DOI: 10.3390/cancers13225848] [Citation(s) in RCA: 31] [Impact Index Per Article: 10.3] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/06/2021] [Accepted: 11/11/2021] [Indexed: 12/24/2022] Open
Abstract
Endometrial cancer (EC) is the most common gynaecologic malignancy in developed countries. The main challenge in EC management is to correctly estimate the risk of metastases at diagnosis and the risk to develop recurrences in the future. Risk stratification determines the need for surgical staging and adjuvant treatment. Detection of occult, microscopic metastases upstages patients, provides important prognostic information and guides adjuvant treatment. The molecular classification subdivides EC into four prognostic subgroups: POLE ultramutated; mismatch repair deficient (MMRd); nonspecific molecular profile (NSMP); and TP53 mutated (p53abn). How surgical staging should be adjusted based on preoperative molecular profiling is currently unknown. Moreover, little is known whether and how other known prognostic biomarkers affect prognosis prediction independent of or in addition to these molecular subgroups. This review summarizes the factors incorporated in surgical staging (i.e., peritoneal washing, lymph node dissection, omentectomy and peritoneal biopsies), and its impact on prognosis and adjuvant treatment decisions in an era of molecular classification of EC. Moreover, the relation between FIGO stage and molecular classification is evaluated including the current gaps in knowledge and future perspectives.
Collapse
Affiliation(s)
- Jenneke C. Kasius
- Department of Obstetrics & Gynaecology, Amsterdam University Medical Centres, 1105 AZ Amsterdam, The Netherlands; (J.C.K.); (J.v.Z.)
| | | | - Kristina Lindemann
- Department of Gynaecologic Oncology, Oslo University Hospital, 0188 Oslo, Norway;
- Institute of Clinical Medicine, University of Oslo, 0318 Oslo, Norway
| | - David Forsse
- Department of Gynaecology and Obstetrics, Haukeland University Hospital, 5021 Bergen, Norway; (D.F.); (C.K.)
| | - Judith van Zwol
- Department of Obstetrics & Gynaecology, Amsterdam University Medical Centres, 1105 AZ Amsterdam, The Netherlands; (J.C.K.); (J.v.Z.)
| | - Gunnar B. Kristensen
- Institute for Cancer Genetics and Informatics, Department of Oncology, Division of Cancer Medicine, Oslo University Hospital, 0424 Oslo, Norway;
| | - Camilla Krakstad
- Department of Gynaecology and Obstetrics, Haukeland University Hospital, 5021 Bergen, Norway; (D.F.); (C.K.)
| | - Henrica M. J. Werner
- Department of Obstetrics and Gynaecology, GROW, Maastricht University School for Oncology & Developmental Biology, 6202 AZ Maastricht, The Netherlands;
| | - Frédéric Amant
- Department of Obstetrics & Gynaecology, Amsterdam University Medical Centres, 1105 AZ Amsterdam, The Netherlands; (J.C.K.); (J.v.Z.)
- Department of Oncology, KU Leuven, 3000 Leuven, Belgium
- Department of Gynaecology, Netherlands Cancer Institute, 1066 CX Amsterdam, The Netherlands
- Correspondence:
| |
Collapse
|